A new erasing method for a single-voltage long-endurance flashmemory

A new method to erase a standard (double-poly, stacked-gate NOR-type) flash cell is proposed. The method, still using the tunneling mechanism to extract electrons from the floating gate, is based on the concept of keeping the electric field constant during the whole erasing operation. The new method has two main advantages with respect to the conventional one: (1) it does not depend on the supply voltage variation and (2) it allows a better reliability in terms of endurance-induced stress. Results have shown that flash device performances are greatly improved in terms of stability and endurance reliability up to one million cycles     

Ultralow-voltage boron-doped diamond field emitter vacuum diode

A boron-doped diamond field emitter diode with ultralow turn-on voltage and high emission current is reported. The diamond field emitter diode structure with a built-in cap was fabricated using molds and electrostatic bonding techniques. The emission current versus anode voltage of the capped diamond emitter diode with boron doping, sp² content, and vacuum thermal electric (VTE) treatment shows a very low turn-on voltage of 2 V. A high emission current of 1 μA at an anode voltage of less than 10 V can be obtained from a single diamond tip. The turn-on voltage is significantly lower than comparable silicon field emitters     

Communication network protocol for real-time distributed controland its LSI implementation

A new token-passing mechanism, priority token passing, which features real-time access and fast detection and recovery of transmission errors, is discussed in detail in comparison with standard token-passing protocols, and its large-scale integration (LSI)-oriented design concept is described. Priority token passing includes only a small performance overhead, due to its switching functions, which can change network topology from ring to broadcast medium. A token-holding node passes the token to another node after determining the successor through priority comparison. Errors occurring during token passing can, thus, be detected and corrected simply and promptly. Priority token passing has a simple hardware implementation, requiring only small additions to the frame control circuitry, and has a small implementation overhead. The priority token-passing protocol and two other important network communication functions, dual ring network reconfiguration and high-level data link control (HDLC) normal response mode-based message transmission, are designed as a single finite-state machine, and implemented into a compact LSI chip. This integrated instrument network (IINET) chip provides complete network communication services and requires only three additional external electronic components for operation     

Evaluation of multicast routing algorithms for real-timecommunication on high-speed networks

Multicast (MC) routing algorithms capable of satisfying the quality of service (QoS) requirements of real-time applications will be essential for future high-speed networks. We compare the performance of all of the important MC routing algorithms when applied to networks with asymmetric link loads. Each algorithm is judged based on the quality of the MC trees it generates and its efficiency in managing the network resources. Simulation results over random networks show that unconstrained algorithms are not capable of fulfilling the QoS requirements of real-time applications in wide-area networks. Simulations also reveal that one of the unconstrained algorithms, reverse path multicasting (RPM), is quite inefficient when applied to asymmetric networks. We study how combining routing with resource reservation and admission control improves the RPM's efficiency in managing the network resources. The performance of one semiconstrained heuristic, MSC, three constrained Steiner tree (CST) heuristics, Kompella, Pasquale, and Polyzos (1992), constrained adaptive ordering (CAO), and bounded shortest multicast algorithm (BSMA), and one constrained shortest path tree (CSPT) heuristic, the constrained Dijkstra heuristic (CDKS) are also studied. Simulations show that the semiconstrained and constrained heuristics are capable of successfully constructing MC trees which satisfy the QoS requirements of real-time traffic. However, the cost performance of the heuristics varies. The BSMA's MC trees are lower in cost than all other constrained heuristics. Finally, we compare the execution times of all algorithms, unconstrained, semiconstrained, and constrained     

Time-resolved synchrotron X-ray study of crystalline phase transition in poly(aryl ether ketone ketone) containing alternated terephthalic/isophthalic moieties

Unique crystallization and melting behavior in poly(aryl ether ketone ketone) containing alternated terephthalic and isophthalic moieties were studied by time-resolved synchrotron x-ray methods. Recently, this material has been shown to exhibit three polymorphs (forms I, II, and III). In this work, we further investigated their distinctive thermal properties and found that form I is the dominating and the most thermally stable phase while form II is favored by fast nucleation conditions and is the least stable phase. On the other hand, form III represents a minor intermediate phase that usually coexists with form I and can be transferred from form II and to form I. Structural and morphological changes in form I have been followed by simultaneous wide-angle x-ray diffraction (WAXD)/small-angle x-ray scattering (SAXS) measurements during cold- or melt-crystallization and subsequent melting. In all cases, a larger dimensional change was found in the crystallographic a-axis than the b-axis during heating and cooling. This may be due to the greater lateral stress variation with respect to temperature along the a direction of the primary lamellae which is induced by either the formation of secondary lamellae or the preferential chain-folding direction in poly(aryl ether ketone ketone)s. During the phase transitions of form II ← III in the cold-crystallized specimen and form III ← I in the melt-crystallized samples, lamellar variables (long period, lamellar thickness, and invariant) obtained from SAXS remain almost constant. This indicates that the density distribution in the long spacing is independent of the melting in form II or III. For melt-crystallization, the corresponding changes in unit-cell dimensions and lamellar morphology during the annealing-induced low endotherm are most consistent with the argument that these changes are due to the melting of thin lamellar population. © 1995 John Wiley & Sons, Inc.     

Parallel, self-organizing, hierarchical neural networks. II

For pt.I see IEEE Trans. Neural Networks, vol.1, p.167-78 (1990). Parallel, self-organizing, hierarchical neural networks (PSHNNs) involve a number of stages with error detection at the end of each stage, i.e., rejection of error-causing vectors, which are then fed into the next stage after a nonlinear transformation. The stages operate in parallel during testing. Statistical properties and the mechanisms of vector rejection of the PSHNN are discussed in comparison to the maximum likelihood method and the backpropagation network. The PSHNN is highly fault tolerant and robust against errors in the weight values due to the adjustment of the error detection bounds to compensate errors in the weight values. These properties are exploited to develop architectures for programmable implementations in which the programmable parts are reduced to on-off or bipolar switching operations for bulk computations and attenuators for pointwise operations     

Computation of confined coflow jets with three turbulence models

A numerical study of confined jets in a cylindrical duct is carried out to examine the performance of two recently proposed turbulence models: an RNG-based K-? model and a realizable Reynolds stress algebraic equation model. The former is of the same form as the standard K-? model but has different model coefficients. The latter uses an explicit quadratic stress-strain relationship to model the turbulent stresses and is capable of ensuring the positivity of each turbulent normal stress. The flow considered involves recirculation with unfixed separation and reatachment points and severe adverse pressure gradients, thereby providing a valuable test of the predictive capability of the models for complex flows. Calculations are performed with a finite volume procedure. Numerical credibility of the solutions is ensured by using second-order-accurate differencing schemes and sufficiently fine grids. Calculations with the standard K-? model are also made for comparison. Detailed comparisons with experiments show that the realizable Reynolds stress algebraic equation model consistently works better than does the standard K-? model in capturing the essential flow features, while the RNG-based K-? model does not seem to give improvements over the standard K-? model under the flow conditions considered.     

Packed hyper-ellipsoid classifiers

Pattern classification is frequently performed using the k-nn algorithm or a neural network. The choice of parameters for the former is often difficult and the amount of data which has to be stored in the classifier can be high. Neural network classifiers can overcome some of these problems but learning is often unreliable and slow. An alternative which combines some of the best features of the k-nn and neural network classifier is described by the authors. The classifier is called a packed hyper-ellipsoid classifier     

Vector algorithm for Reed-Muller expansions

A novel vector algorithm for Reed-Muller (RM) expansions is proposed which can save as many as a factor of 2^n-1 memory elements compared with previous matrix algorithms